{"id":{"repo_id":"purdue-thes","oai_identifier":"oai:docs.lib.purdue.edu:open_access_dissertations-2546"},"canonical_url":"https://search.dev.ndltd.org/etd/purdue-thes/oai:docs.lib.purdue.edu:open_access_dissertations-2546","repository":{"repo_id":"purdue-thes","name":"Purdue University","base_url":"https://docs.lib.purdue.edu/do/oai/"},"display":{"title":"Advanced Statistical Modeling for Model-Based Iterative Reconstruction for Single-Energy and Dual-Energy X-Ray CT","abstract":"Model-based iterative reconstruction (MBIR) has been increasingly broadly applied as an improvement over traditional, analytical image reconstruction methods in X-ray CT, primarily due to its significant advantage in drastic dose reduction without diagnostic loss. Early success of the method in conventional CT has encouraged the extension to a wide range of applications that includes more advanced imaging modalities, such as dual-energy X-ray CT, and more challenging imaging conditions, such as low-dose and sparse-sampling scans, each requiring refined statistical models including the data model and the prior model. In this dissertation, we developed an MBIR algorithm for dual-energy CT that included a joint data-likelihood model to account for correlated data noise. Moreover, we developed a Gaussian-Mixture Markov random filed (GM-MRF) image model that can be used as a very expressive prior model in MBIR for X-ray CT reconstruction. The GM-MRF model is formed by merging individual patch-based Gaussian-mixture models and therefore leads to an expressive MRF model with easily estimated parameters. Experimental results with phantom and clinical datasets have demonstrated the improvement in image quality due to the advanced statistical modeling.","abstract_html":"Model-based iterative reconstruction (MBIR) has been increasingly broadly applied as an improvement over traditional, analytical image reconstruction methods in X-ray CT, primarily due to its significant advantage in drastic dose reduction without diagnostic loss. Early success of the method in conventional CT has encouraged the extension to a wide range of applications that includes more advanced imaging modalities, such as dual-energy X-ray CT, and more challenging imaging conditions, such as low-dose and sparse-sampling scans, each requiring refined statistical models including the data model and the prior model. In this dissertation, we developed an MBIR algorithm for dual-energy CT that included a joint data-likelihood model to account for correlated data noise. Moreover, we developed a Gaussian-Mixture Markov random filed (GM-MRF) image model that can be used as a very expressive prior model in MBIR for X-ray CT reconstruction. The GM-MRF model is formed by merging individual patch-based Gaussian-mixture models and therefore leads to an expressive MRF model with easily estimated parameters. Experimental results with phantom and clinical datasets have demonstrated the improvement in image quality due to the advanced statistical modeling.","abstract_has_math":false,"creators":["Zhang, Ruoqiao"],"institution":null,"degree_name":"Doctor of Philosophy (PhD)","degree_level":"Dissertation","degree_discipline":"Electrical and Computer Engineering","degree_department":null,"school":null,"contributors":["Charles Bouman","Mary Comer","Thomas Talavage","Ken Sauer","Jean-Baptiste Thibault"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2015,"date_issued":"2015-01-01T08:00:00Z","date_published":"2015-01-01T08:00:00Z","updated_at":"2026-07-24T03:54:38Z","subjects":["Computed tomography","Model-based iterative reconstruction","Prior modeling","Statistical modeling"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://docs.lib.purdue.edu/open_access_dissertations/1330","outbound_label":"Repository record","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Charles Bouman","Mary Comer","Thomas Talavage","Ken Sauer","Jean-Baptiste Thibault"]},{"key":"dc:creator","label":"Author","values":["Zhang, Ruoqiao"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"thesis:degree_discipline","label":"Discipline","values":["Electrical and Computer Engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Doctor of Philosophy (PhD)"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Computed tomography","Model-based iterative reconstruction","Prior modeling","Statistical modeling"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://docs.lib.purdue.edu/open_access_dissertations/1330"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Model-based iterative reconstruction (MBIR) has been increasingly broadly applied as an improvement over traditional, analytical image reconstruction methods in X-ray CT, primarily due to its significant advantage in drastic dose reduction without diagnostic loss. Early success of the method in conventional CT has encouraged the extension to a wide range of applications that includes more advanced imaging modalities, such as dual-energy X-ray CT, and more challenging imaging conditions, such as low-dose and sparse-sampling scans, each requiring refined statistical models including the data model and the prior model. In this dissertation, we developed an MBIR algorithm for dual-energy CT that included a joint data-likelihood model to account for correlated data noise. Moreover, we developed a Gaussian-Mixture Markov random filed (GM-MRF) image model that can be used as a very expressive prior model in MBIR for X-ray CT reconstruction. The GM-MRF model is formed by merging individual patch-based Gaussian-mixture models and therefore leads to an expressive MRF model with easily estimated parameters. Experimental results with phantom and clinical datasets have demonstrated the improvement in image quality due to the advanced statistical modeling."]},{"key":"dc:title","label":"Title","values":["Advanced Statistical Modeling for Model-Based Iterative Reconstruction for Single-Energy and Dual-Energy X-Ray CT"]}]}],"canonical_facts":{"dc:contributor":["Charles Bouman","Mary Comer","Thomas Talavage","Ken Sauer","Jean-Baptiste Thibault"],"dc:creator":["Zhang, Ruoqiao"],"dc:description.abstract":["Model-based iterative reconstruction (MBIR) has been increasingly broadly applied as an improvement over traditional, analytical image reconstruction methods in X-ray CT, primarily due to its significant advantage in drastic dose reduction without diagnostic loss. Early success of the method in conventional CT has encouraged the extension to a wide range of applications that includes more advanced imaging modalities, such as dual-energy X-ray CT, and more challenging imaging conditions, such as low-dose and sparse-sampling scans, each requiring refined statistical models including the data model and the prior model. In this dissertation, we developed an MBIR algorithm for dual-energy CT that included a joint data-likelihood model to account for correlated data noise. Moreover, we developed a Gaussian-Mixture Markov random filed (GM-MRF) image model that can be used as a very expressive prior model in MBIR for X-ray CT reconstruction. The GM-MRF model is formed by merging individual patch-based Gaussian-mixture models and therefore leads to an expressive MRF model with easily estimated parameters. Experimental results with phantom and clinical datasets have demonstrated the improvement in image quality due to the advanced statistical modeling."],"dc:identifier":["https://docs.lib.purdue.edu/open_access_dissertations/1330"],"dc:subject":["Computed tomography","Model-based iterative reconstruction","Prior modeling","Statistical modeling"],"dc:title":["Advanced Statistical Modeling for Model-Based Iterative Reconstruction for Single-Energy and Dual-Energy X-Ray CT"],"thesis:degree_discipline":["Electrical and Computer Engineering"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Doctor of Philosophy (PhD)"]},"updated_at":"2026-07-24T03:54:38Z"}